5 5.1.2alpha (2012-07-04)
7 * All fixes from 5.0.3 and 5.0.4
11 - Fixed a deadlock and an invalid free() in the threaded encoder.
13 - Added support for symbol versioning. It is enabled by default
14 on GNU/Linux, other GNU-based systems, and FreeBSD.
16 - Use SHA-256 implementation from the operating system if one is
17 available in libc, libmd, or libutil. liblzma won't use e.g.
18 OpenSSL or libgcrypt to avoid introducing new dependencies.
20 - Fixed liblzma.pc for static linking.
22 - Fixed a few portability bugs.
24 * xz --decompress --single-stream now fixes the input position after
25 successful decompression. Now the following works:
27 echo foo | xz > foo.xz
28 echo bar | xz >> foo.xz
29 ( xz -dc --single-stream ; xz -dc --single-stream ) < foo.xz
31 Note that it doesn't work if the input is not seekable
32 or if there is Stream Padding between the concatenated
35 * xz -lvv now shows the minimum xz version that is required to
36 decompress the file. Currently it is 5.0.0 for all supported .xz
37 files except files with empty LZMA2 streams require 5.0.2.
39 * Added an *incomplete* implementation of --block-list=SIZES to xz.
40 It only works correctly in single-threaded mode and when
41 --block-size isn't used at the same time. --block-list allows
42 specifying the sizes of Blocks which can be useful e.g. when
43 creating files for random-access reading.
46 5.1.1alpha (2011-04-12)
48 * All fixes from 5.0.2
50 * liblzma fixes that will also be included in 5.0.3:
52 - A memory leak was fixed.
54 - lzma_stream_buffer_encode() no longer creates an empty .xz
55 Block if encoding an empty buffer. Such an empty Block with
56 LZMA2 data would trigger a bug in 5.0.1 and older (see the
57 first bullet point in 5.0.2 notes). When releasing 5.0.2,
58 I thought that no encoder creates this kind of files but
61 - Validate function arguments better in a few functions. Most
62 importantly, specifying an unsupported integrity check to
63 lzma_stream_buffer_encode() no longer creates a corrupt .xz
64 file. Probably no application tries to do that, so this
65 shouldn't be a big problem in practice.
67 - Document that lzma_block_buffer_encode(),
68 lzma_easy_buffer_encode(), lzma_stream_encoder(), and
69 lzma_stream_buffer_encode() may return LZMA_UNSUPPORTED_CHECK.
71 - The return values of the _memusage() functions are now
74 * Support for multithreaded compression was added using the simplest
75 method, which splits the input data into blocks and compresses
76 them independently. Other methods will be added in the future.
77 The current method has room for improvement, e.g. it is possible
78 to reduce the memory usage.
80 * Added the options --single-stream and --block-size=SIZE to xz.
82 * xzdiff and xzgrep now support .lzo files if lzop is installed.
83 The .tzo suffix is also recognized as a shorthand for .tar.lzo.
85 * Support for short 8.3 filenames under DOS was added to xz. It is
86 experimental and may change before it gets into a stable release.
93 - Fix lzma_index_init(). It could crash if memory allocation
96 - Fix the possibility of an incorrect LZMA_BUF_ERROR when a BCJ
97 filter is used and the application only provides exactly as
98 much output space as is the uncompressed size of the file.
100 - Fix a bug in doc/examples_old/xz_pipe_decompress.c. It didn't
101 check if the last call to lzma_code() really returned
102 LZMA_STREAM_END, which made the program think that truncated
105 - New example programs in doc/examples (old programs are now in
106 doc/examples_old). These have more comments and more detailed
109 * Fix "xz -lvv foo.xz". It could crash on some corrupted files.
111 * Fix output of "xz --robot -lv" and "xz --robot -lvv" which
112 incorrectly printed the filename also in the "foo (x/x)" format.
114 * Fix exit status of "xzdiff foo.xz bar.xz".
116 * Fix exit status of "xzgrep foo binary_file".
118 * Fix portability to EBCDIC systems.
120 * Fix a configure issue on AIX with the XL C compiler. See INSTALL
123 * Update French, German, Italian, and Polish translations.
130 - A memory leak was fixed.
132 - lzma_stream_buffer_encode() no longer creates an empty .xz
133 Block if encoding an empty buffer. Such an empty Block with
134 LZMA2 data would trigger a bug in 5.0.1 and older (see the
135 first bullet point in 5.0.2 notes). When releasing 5.0.2,
136 I thought that no encoder creates this kind of files but
139 - Validate function arguments better in a few functions. Most
140 importantly, specifying an unsupported integrity check to
141 lzma_stream_buffer_encode() no longer creates a corrupt .xz
142 file. Probably no application tries to do that, so this
143 shouldn't be a big problem in practice.
145 - Document that lzma_block_buffer_encode(),
146 lzma_easy_buffer_encode(), lzma_stream_encoder(), and
147 lzma_stream_buffer_encode() may return LZMA_UNSUPPORTED_CHECK.
149 - The return values of the _memusage() functions are now
152 * Fix command name detection in xzgrep. xzegrep and xzfgrep now
153 correctly use egrep and fgrep instead of grep.
155 * French translation was added.
160 * LZMA2 decompressor now correctly accepts LZMA2 streams with no
161 uncompressed data. Previously it considered them corrupt. The
162 bug can affect applications that use raw LZMA2 streams. It is
163 very unlikely to affect .xz files because no compressor creates
164 .xz files with empty LZMA2 streams. (Empty .xz files are a
165 different thing than empty LZMA2 streams.)
167 * "xz --suffix=.foo filename.foo" now refuses to compress the
168 file due to it already having the suffix .foo. It was already
169 documented on the man page, but the code lacked the test.
171 * "xzgrep -l foo bar.xz" works now.
173 * Polish translation was added.
178 * xz --force now (de)compresses files that have setuid, setgid,
179 or sticky bit set and files that have multiple hard links.
180 The man page had it documented this way already, but the code
183 * gzip and bzip2 support in xzdiff was fixed.
187 * Minor fix to Czech translation
192 Only the most important changes compared to 4.999.9beta are listed
193 here. One change is especially important:
195 * The memory usage limit is now disabled by default. Some scripts
196 written before this change may have used --memory=max on xz command
197 line or in XZ_OPT. THESE USES OF --memory=max SHOULD BE REMOVED
198 NOW, because they interfere with user's ability to set the memory
199 usage limit himself. If user-specified limit causes problems to
200 your script, blame the user.
202 Other significant changes:
204 * Added support for XZ_DEFAULTS environment variable. This variable
205 allows users to set default options for xz, e.g. default memory
206 usage limit or default compression level. Scripts that use xz
207 must never set or unset XZ_DEFAULTS. Scripts should use XZ_OPT
208 instead if they need a way to pass options to xz via an
209 environment variable.
211 * The compression settings associated with the preset levels
212 -0 ... -9 have been changed. --extreme was changed a little too.
213 It is now less likely to make compression worse, but with some
214 files the new --extreme may compress slightly worse than the old
217 * If a preset level (-0 ... -9) is specified after a custom filter
218 chain options have been used (e.g. --lzma2), the custom filter
219 chain will be forgotten. Earlier the preset options were
220 completely ignored after custom filter chain options had been
223 * xz will create sparse files when decompressing if the uncompressed
224 data contains long sequences of binary zeros. This is done even
225 when writing to standard output that is connected to a regular
226 file and certain additional conditions are met to make it safe.
228 * Support for "xz --list" was added. Combine with --verbose or
229 --verbose --verbose (-vv) for detailed output.
231 * I had hoped that liblzma API would have been stable after
232 4.999.9beta, but there have been a couple of changes in the
233 advanced features, which don't affect most applications:
235 - Index handling code was revised. If you were using the old
236 API, you will get a compiler error (so it's easy to notice).
238 - A subtle but important change was made to the Block handling
239 API. lzma_block.version has to be initialized even for
240 lzma_block_header_decode(). Code that doesn't do it will work
241 for now, but might break in the future, which makes this API
244 * The major soname has been bumped to 5.0.0. liblzma API and ABI
245 are now stable, so the need to recompile programs linking against
246 liblzma shouldn't arise soon.